[Spark/Hive] -- Spark UDF

1、定义UDF

import java.util.Random

def randomPrefixUDF(value: String): String = {
  new Random().nextInt(10).toString() + "_" + value.toInt
}

def removePrefixUDF(value: String): String = {
  value.split("_")(1)
}

2、注册UDF

spark.udf.register("random_prefix", (value:String) => randomPrefixUDF(value))
spark.udf.register("remove_random_prefix", (value:String) => removePrefixUDF(value))

3、使用示例

select random_prefix(user_id), remove_random_prefix(random_prefix(user_id))
from tb_user
limit 10